Skip to content

usage-guide.md

Latest commit

 

History

History
78 lines (51 loc) · 2.1 KB

usage-guide.md

File metadata and controls

78 lines (51 loc) · 2.1 KB
Raw

Differential miRNA Expression - Usage Guide

Overview

Identify differentially expressed miRNAs between conditions using DESeq2 or edgeR with considerations specific to small RNA data.

Prerequisites

BiocManager::install(c('DESeq2', 'edgeR', 'apeglm', 'EnhancedVolcano', 'pheatmap'))

Quick Start

Tell your AI agent:

  • "Find differentially expressed miRNAs between treatment and control"
  • "Run DESeq2 on my miRNA count matrix"
  • "Create a volcano plot of DE miRNAs"
  • "Filter for significant miRNAs with |log2FC| > 1"

Example Prompts

Basic DE Analysis

"Run differential expression on my miRge3 counts"

"Compare miRNA expression between tumor and normal samples"

"Find miRNAs with adjusted p-value < 0.05"

Visualization

"Create a volcano plot of differentially expressed miRNAs"

"Make a heatmap of significant miRNAs"

"Plot MA of miRNA expression changes"

Export and Filtering

"Export significant miRNAs to CSV"

"Filter for miRNAs with at least 2-fold change"

"List the top 20 upregulated miRNAs"

What the Agent Will Do

  1. Load miRNA count matrix and sample metadata
  2. Create DESeq2 dataset with appropriate design
  3. Filter low-expressed miRNAs (< 10 total reads)
  4. Run DESeq2 and apply apeglm shrinkage
  5. Filter by significance (padj < 0.05, |log2FC| > 1)
  6. Generate visualizations and export results

Tips

  • Filter low counts - miRNAs with < 10 total reads are unreliable
  • Use apeglm shrinkage - improves log2FC estimates for low-count miRNAs
  • Check normalization - miRNA libraries can have different compositions
  • Multiple testing - always use adjusted p-values
  • |log2FC| > 1 is standard for biologically meaningful changes